Product-led growth (PLG)

Product-led growth (PLG) is a go-to-market strategy where the product is the primary acquisition, activation, and expansion channel. Users experience value before they talk to sales or see a pricing page. The mechanism: make the product fast enough to value that a free user hits a usage threshold (a Product Qualified Lead, or PQL) that predicts conversion or expansion. PLG was coined by Blake Bartlett at OpenView Partners in 2016 and popularized by Wes Bush’s 2019 book. By 2025, 58% of B2B SaaS companies use some form of PLG motion, versus a niche strategy in 2018.

The weak interview answer treats PLG as a synonym for “viral” or “no sales team.” Slack is the canonical example, but Slack grew through network effects, not pure PLG. PLG is about the product delivering value fast enough to convert free users to paid, with or without virality. Interviewers at Figma, Notion, Atlassian, or Cursor hear the viral shorthand and know the candidate has read a blog post but has not thought operationally about growth.

The core mechanics

Three things have to work together for PLG to function:

Activation. The user reaches a defined first-value moment before hitting friction or a paywall. Only 34% of PLG companies actively track activation rates despite it being the single strongest predictor of free-to-paid conversion. If you cannot name your activation event and the window, you are not running PLG operationally.

The PQL. A Product Qualified Lead is a free user who has crossed a defined usage threshold that signals readiness to convert or expand. Not a demographic signal, not a marketing funnel stage: a behavioral signal from within the product. Examples: invited 3 or more teammates, created 5 or more assets, used the core action 3 times within 7 days. The PQL definition is the artifact that connects the product team to sales. When a free user hits it, an outbound motion (automated or human) can begin without disrupting users who are not ready.

Expansion via NRR. PLG monetizes twice: acquisition (free-to-paid conversion) and expansion (seat expansion, tier upgrades, usage overages). Net Revenue Retention is the health metric for the second loop. Figma’s PLG worked because any designer could open a file without an account, which dragged in the whole org, which then converted through a company-wide contract. The viral invite was the acquisition; the enterprise contract was the expansion. Both are PLG.

When PLG works and when it does not

PLG works when: the product delivers perceivable value to an individual user before requiring organizational buy-in, the core loop is short enough to reach the aha moment in hours or days rather than weeks, and the COGS per free user is low enough to sustain a generous trial without destroying unit economics.

PLG does not work when: compliance or security requires pre-sale approval (regulated industries, government, healthcare), the product requires significant behavior change before delivering value (implementation-heavy ERP, for example), or COGS per free user exceeds your acceptable CAC budget. That last condition is new and important in 2026.

What has changed in 2026

The classic PLG playbook assumed that free tiers could be indefinitely generous because the marginal cost of serving another user was near zero. That assumption breaks for AI-native products where each query carries real inference cost. Cursor reached $500M ARR in under 24 months and $200M ARR before hiring its first enterprise sales rep, but Cursor can run this because its free tier is carefully constrained. Lovable reached $100M ARR in 8 months. Menlo Ventures data shows 27% of AI application spend flows through PLG motions versus 7% for traditional SaaS.

The response from most AI-native products: shift from indefinitely generous free tiers to time-boxed trials (7 to 14 days), credit systems, and hard usage caps. Agentic onboarding (AI-configured setup, no human working through a checklist) lifts free-to-paid conversion from the typical 3 to 8% range to 25 to 30%. And Netlify reports that 80% of new signups are now agents, not humans, which means activation metrics, onboarding flows, and PQL definitions built for human users are already stale for parts of the market.

The 2026 reframe: feasibility is essentially free. PLG’s job is no longer to bypass a sales team; it is to compress time-to-lovable-value fast enough that users form a habit before a competitor even runs a demo. Viable (real market, unit economics that work) and lovable (the product earns trust through the experience itself) are the two tests PLG has to pass. Usable is the floor, not the ceiling.

PLG vs. product-led sales

Pure PLG (no sales) is rare above $10M ARR. The dominant model at Figma, Notion, and Cursor is product-led sales: self-serve adoption first, then a sales layer triggered by PQL signals. The sales rep enters only when behavioral data says the account is ready, which means the first call is not discovery, it is expansion. This is why PQL definition is a cross-functional artifact, not a product-only decision.

Metrics to know

Activation rate (did the user reach first value within the defined window)
Time-to-value (TTV): median time from signup to activation event
PQL conversion rate: free users who crossed the PQL threshold and converted to paid
Expansion MRR and NRR: the health of the second PLG loop
Viral coefficient: relevant if the product has a built-in sharing or invite mechanic, but not required for PLG

Weak vs. strong interview answer

weak

"PLG is when your product grows itself, like Slack or Dropbox. Instead of having a sales team, the product spreads virally and users sign up on their own." This describes the outcome, not the mechanism. It conflates viral growth with PLG. It offers no judgment about when PLG works or fails. It ignores the 2026 context where free tiers are breaking down and product-led sales is the dominant hybrid model. An interviewer at a PLG-native company hears this and knows the candidate has not thought operationally about growth.

strong

"PLG is a GTM strategy where the product itself is the primary acquisition, activation, and expansion channel. Users experience value before they talk to sales or see a pricing page. The mechanism: make the product fast enough to value that a free user hits a usage threshold, what we call a PQL, that predicts they will convert or expand. Figma is the case I find most instructive: anyone on a design team could open a file without an account, which dragged in the whole org, then the org converted through an enterprise contract. The key metrics are activation rate, time-to-value, PQL conversion rate, and NRR, because in a PLG model expansion from existing users is as important as new acquisition. In 2026 I would add a nuance: pure PLG is under pressure for AI-native products because inference costs mean indefinitely free is not viable. The winning motion is product-led sales: self-serve adoption first, then a sales layer triggered by PQL signals, so the economics actually work."

For how PLG fits into the full growth loop, see activation, CAC, and AARRR. If you are interviewing for a growth-focused role, growth PM covers how this shows up in role-specific interview rounds.